Turnitin updates tool to detect AI bypassers & humanised text

Turnitin has announced an update to its AI writing detection tools, introducing new capabilities aimed at identifying AI-generated content that has been intentionally modified to evade detection.

The surge in use of AI "bypasser" tools, sometimes referred to as "humanisers", has presented fresh challenges for educators and academic institutions over the last two years. These tools allow students to mask content generated by large language models (LLMs) to resemble human writing, complicating the process of verifying the authenticity and originality of student submissions.

Turnitin's latest feature provides educators with the ability to identify both standard AI-generated writing and work that has been subsequently altered by humaniser software. The new capability is integrated directly into Turnitin's existing workflow, reducing reliance on third-party detection tools.

Addressing humanised AI content

According to Turnitin, bypasser tools have raised concerns because they make it significantly easier for students not only to use AI for written assignments, but to actively conceal its involvement. As a result, questions over academic integrity and the validity of learning outcomes have intensified in educational circles.

James Thorley, Regional Vice President for Asia-Pacific at Turnitin, emphasised the integrated nature of the new service. He said, "Turnitin's new AI bypasser feature, effortlessly blends into Turnitin's workflow. Here educators are now equipped with tools to detect AI-generated content and transform potential misuse into a learning opportunity for students."

Turnitin's latest AI bypasser detection does not require any separate integration or product and aims to minimise security and data privacy concerns related to using additional, external services. When enabled, the system flags both original AI-generated text and that which has been disguised by bypasser tools.

Annie Chechitelli, Chief Product Officer at Turnitin, provided additional context on the issue of AI misuse in education. She stated,

"While cheating has always been a concern of educators, with the rise of AI, there has been a new category formed of cheating providers. These companies exist to profit from students' misuse of AI by providing free and easy access to humanisers to conceal AI-generated content," said Annie Chechitelli, chief product officer, Turnitin. "In service of our ongoing commitment to AI writing transparency, we've updated our software to detect leading AI bypasser modifications so educators can uphold academic integrity against this emerging challenge."

Concerns over academic integrity

The availability and increasing sophistication of AI bypasser tools have prompted concerns about their effect on the educational system. By allowing students to present work that may not be their own, such tools are seen as undermining opportunities for genuine learning and eroding the trust between educators and students.

AI-driven tools such as Turnitin Clarity have demonstrated positive potential to enhance instruction and assessment. However, the misuse of such technologies to commit academic dishonesty remains a growing threat within the sector. The company's new bypasser detection feature aims to address these risks by helping educators more confidently confirm the originality of submitted work.

In Australia, the Tertiary Education Quality and Standards Agency (TEQSA) has taken steps to promote assessment reform and uphold academic integrity with the launch of its Gen AI Knowledge Hub. However, challenges persist, including the widespread use of humaniser tools and inconsistent access to educational guidance on AI.

Turnitin stated that its new feature is intended to give educators "tools to detect AI-generated content and transform potential misuse into a learning opportunity for students." The company sees the development as an extension of its commitment to support academic institutions in fostering an environment of honesty and trust.

Product availability

The AI bypass detection capability is now available for customers using Turnitin's Originality add-on and for iThenticate 2.0 customers who have access to additional AI functionality. The detection of bypasser-modified content is currently supported for English-language submissions.

This update follows the recent launch of Turnitin Clarity and an overhauled version of Turnitin Feedback Studio, indicating a continued focus by the company on adapting its services to developments in AI and large language models. The company noted its intention to assist educators and administrators in upholding standards of academic integrity as new technologies emerge.